Horsing around

If there is one thing which I really missed out on and regret about my PhD, it is that I was unable to truly collaborate with the other members of the perissodactyl research group within the FunMorph Lab. My PhD supervisor, dr. Sandra Nauwelaerts, was (and remains) an expert in equine locomotion and biomechanics. Sandra’s other student, dr. Mariëlle Kaashoek, wrote her PhD on the joint kinematics of equids, with some overlap which included tapirs but nothing which could really be called an “opportunity ” for genuine collaboration. And Sandra’s promoter (and now my post-doctoral supervisor) Professor dr. Peter Aerts, who is a world renowned biomechanist and functional morphologist, was involved in my project whenever we needed biomechanical clarification and insight, but essentially allowed Sandra to supervise me autonomously. It all worked – there was no need to rock the boat – but on seeing the interplay and collaboration between other members of the FunMorph lab (the “lizard people” as most of them are affectionately called [by me and others]), I couldn’t help but think on what might have been…

And so, it is with great pride and enjoyment that I can announce that the collaboration woes have been vanquished! Throughout the rather disappointing years of 2020-2021 (virus, masks, no colleagues, no holidays, etc.), we have been working on a book chapter for a volume on equids published by Springer, covering their ecology, evolution, behaviour, human interactions – essentially everything there is to know about equids! Here we can give you all a sneak peek at our book chapter: Evolution of the Equid Limb.

“The Marsh Series Revisited”, adpated from Kaashoek & MacLaren et al. (2023). Here we see a composite cladogram depicting the major changes in locomotor anatomy through equid evolution, with some very specific and important changes also highlighted (pad-foot to spring-foot; tridactyly to monodactyly). Time is represented along the left of the figure, with numbers representing millions of years before present.

Anyone who knows anything about horse evolution will have come across the “Marsh Series” at some point. They may not have recognised it as such, and perhaps it was not specifically called that, but it will be there in a lot of Biology textbooks for 1st year undergraduates. It was in mine for sure! For me, there is a lot to like about the Marsh series – yes, it is phylogenetically inaccurate in the face of modern cladistic scrutiny; yes, it is a painfully simplistic version of a much more intricate and divergent locomotor transition; but at its heart, no – it’s not actually anatomically inaccurate! Horses (or as I will now more accurately refer to them, equids) underwent an anatomical transition from possessing four toes on their front limbs (as modern tapirs do – yay, tapirs!), through various stages of digit reduction and limb musculoskeletal specialisation throughout the majority of the Cenozoic, with the remaining equids surviving today (all in the genus Equus) possessing a monodactyl (one-toed) foot on both fore- and hind limbs. Now, there are a number of things to factor in here before we delve into how our book chapter deals with this evolutionary transition.

Thing No.1: modern equids are monodactyl, and their genus (Equus) has been around for around 4 million years (Ma). There were also other monodactyl species of equid before Equus (and alongside them) for many millions of years – so Equus was not the first genus to become functionally monodactyl on all four feet. Frankly, equids themselves are not even first mammals to do so – that title goes to the proterotheriids, a group of South American Native Ungulates (SANUs), who achieved this feat in the early Miocene (~20Ma). Anyway, the point is that modern Equus species like donkeys and zebras are the survivors of a monodactyl legacy twice as ancient as hominid bipedalism (i.e. they had one toe before we were walking upright!), and at the end of the day they are not unique in having a single functional toe on each foot.

Thing No.2: Of all the genera of equids that have ever existed, over 80% of those currently described were tridactyl – that is to say, that 4 out of 5 equids that have ever existed had three toes on each foot, not one. Yes – it is the monodactyl equids (and the modern horse most especially) who are the evolutionary oddballs!

And finally, Thing No.3: we did not write our chapter on the evolution of the equid limb to demonstrate that any one group of equids are better or more successful than any others…in fact our base assumption is that through the millennia, equids have been well adapted for their respective habitats and locomotor needs – if you think about it, if they had not been well adapted then (by the rules of Darwinian evolution) the “fittest” they would not have been, and “survived” they would have not! With this key notion in mind – i.e. that equids have never truly been maladapted for their environment – I will endeavour to break down our major argument for the evolution of monodactyly in equids (and, come to think of it, the evolution of monodactyly in any group). We call it the Equal Strength Synthesis, and I will now explain why.

Without delving too much into the physics of movement, here is a brief summary of what you need to know to follow our concept. I use the term ‘concept’ quite deliberately – we base our argument on an idealised condition to make things a bit easier to follow. All the logic behind our synthesis can be found in the chapter (link here), and is rooted in well established biomechanical theory. We use a working example of a “monodactyl Mesohippus” to demonstrate the benefits for an organism (under certain selective pressures) to reduce their number of toes and, in the case of equids, become monodactyl.

Stylised example of a tridactyl Mesohippus (left) and a hypothetical “monodactyl Mesohippus” (right).
Artwork by J. MacLaren, after original painting of Hypohippus by Heinrich Harder

Our concept is very much an update and synthesis of two well-established ideas surrounding horse limb evolution: the Body Mass Thesis and the Locomotor Efficiency Thesis. Broadly speaking, the Body Mass Thesis dictates that a single cylindrical limb bone (e.g. a metapodial) of mass X will be stronger than two or more limb bones of total mass also X. Essentially, having one single metapodial makes the distal limb stronger in compression during loading (e.g. while walking, running or jumping). The Locomotor Efficiency Thesis suggests that the reduction of the distal limb mass (by reducing the limb to a single digit) and the adoption of a highly efficient elastic recoil mechanism (the “spring-foot” apparatus) combined to enable horses to move around at low and medium speeds with high energetic efficiency. Our Equal Strength Synthesis posits that horses have always been strong enough to support their limbs during performance locomotion (e.g. running to evade predators, jumping over obstacles etc.) – so our calculations start off by assuming a STRENGTH of X, rather than a MASS of X.

Here we put together an applied example of the Equal Strength Synthesis. On the left, the three simple cylinders represent a tridactyl horse limb, with side-digits filled in (based on observations of tridactyl equid metapodials). Two simple monodactyl limbs are then presented, one based on the ‘equal mass concept’, and the other based on the ‘equal strength concept’ . We then calculated out the masses and stresses for the simple horse limbs, showing that a monodactyl digit with the same STRENGTH as a tridactyl limb will in fact have up to a 35% reduction in mass!

When the calculations pan out, our simplified single distal limb bone with equal strength as the combined strength of three distal limb bones turned out to have ~30% less mass while maintaining the same length and safety factor! That 30% reduction in mass would significantly benefit the efficiency of locomotion by reducing distal inertia, but also would enable the animal to elongate the limb bone to increase stride length and still be just as strong in compression (though less resistant to bending – there always remains a tradeoff). As a result, our Equal Strength Synthesis represents a biomechanically rooted solution which may explain how equids were able to remain so very successful through such a radical morphological transition. Rather than living on the “edge of failure”, tridactyl equids moved around their respective ecosystems with limb morphologies which had been selected for maximum strength while expending minimal energy during both ordinary and performance locomotion.

With changing ecosystems and associated shifts in diet, body size and ranging behaviour, the strength requirements for the distal limb of equids also shifted, and a single digit became more favourable. The mesaxonic limb of perissodactyls (including horses) made this extreme digit reduction possible – that’s one reason why we don’t see monodactyl members of non-mesaxonic ungulate groups such as giraffes, antelopes or deer today! That said, the same equal strength synthesis concept applies to a multitude of animal groups, including (but not limited to) artiodactyls, notoungulates, macropods, and even ratites! We hope our strength-based approach will prompt discussion and promote greater understanding of digit reduction through time in multiple groups, including iconic taxa such as horses.

Miocene horses from “Doll Ponies” exhibit © Corbin Rainbolt

The book we contributed to “The Equids: A Suite of Splendid Species” is available now from Springer Nature Fascinating Life Sciences series. Our chapter is Chapter 13: “Evolution of the Equid Limb“. We hope you all find it interesting, and don’t forget to check out all the other great work done in this must-read volume on horses, their relatives, their biology and their evolution.

Palaeotheoryum Presents: All the better to eat you with?

Time for something different! As any regular readers of this very irregular blog will be aware, I mostly post about my own studies and make an effort to break down the more complex scientific jargon (which I insist on using in papers) and make it more accessible for non-scientists. Feel free to post a review on how well (or poorly) I am managing!
Anyway, recently I have been contributing to some really interesting studies – particularly coming out of the University of Liège ‘Evolution and Diversity Dynamics Lab’. I assisted with several of these studies mostly at the scanning and writing phases, but I want to share the work of these excellent scientists with you all through their own words. Thus, in a break with the ordinary Palaeotheoryum protocol, here is my first attempt at a review/interview for my colleague’s work! Palaeotheoryum Presents: teeth!

The first study I am going to showcase is hot off the press, published in Proceedings of the Royal Society B: Biological Sciences in September 2022! This paper, part of a multinational endeavour with authors from six countries scattered across the world, brings together morphology, function, ecology, and 3D digitisation to investigate tooth shape and feeding strategy in marine amniotes (mammals and reptiles). Coordinated by the EDDy Lab, the study specifically investigates the variety of shapes, sizes and surface complexities of marine amniote teeth using high resolution 3D scanning and high density geometric morphometric tools, combined with gut contents and other feeding behavioural proxies. I caught up with lead author Professor Valentin Fischer to ask him about the project.

Professor Valentin Fischer using his great height to scan the tricky bits of a mosasaur skull in the display halls of the Royal Belgian Institute of Natural Sciences, Brussels. Image by Jamie MacLaren.

First up – let’s get some context for this study. Within the broader EDDy Lab umbrella, this 2022 study encompasses digitisation methods, pioneers analytical techniques, and focuses on study organisms central to the lab’s recent projects: 3D scanning and geometric morphometrics of marine amniotes. Now, that’s all very well and good, but why teeth? “Teeth are my favorite part of the skeleton!” Valentin elaborates – maybe he is secretly a mammal-worker after all! “Despite their simplicity – in many marine amniotes, at least – they encapsulate so many different signals.” The utility of teeth as a tool used by marine organisms to capture, process, and interact with the surrounding environment, both biotic (e.g. puncturing food) and abiotic (e.g. straining water), makes them fascinating subjects for large scale ecological study. However, the capability for EDDy Lab researchers to capture the full range of tooth variation exhibited by marine amniotes should not be overlooked. “Thanks to the high-precision surface scanners in my lab (both laser and structured-light based)” Valentin continues, “we have accumulated a huge collection of 3D models of marine amniote skulls, jaws…and teeth!“

The study was in part inspired by the qualitative works of Judy Massare, whose seminal research assigned feeding guilds and inferred behaviours about extinct species, especially marine reptiles. The works mostly focused on the features of tooth shape, including the presence of ridges and small serrations. “We explicitly analyse tooth size and orientation, in addition to shape“, Valentin elaborates with growing excitement, “and we also provide a new, repeatable, and quantitative protocol to infer feeding palaeoecology“. The protocol Valentin refers to incorporates a semi-automated landmark-based shape assessment, plus an additional “topographic complexity” component (i.e. we looked at overall shape AND the details on the tooth surface). This procedure is detailed below, a figure used in the article itself!

Brief flow-diagram of the protocol as performed on marine reptile dentition, adapted from Figure 1.

It would be fair to say that the project was not without its challenges, especially given the scope of the sample collected. When asked what the most challenging aspect of the study was, Valentin’s answer was hardly surprising: “Getting enough 3D models during COVID, really!” – as a contributor to that large sample of teeth I can confirm that the biggest headache was specimen availability and accessibility, and a big thank you must go out to the museum collection managers, researchers and collectors who helped us achieve our impressive and highly informative sample. The 3D models of the teeth used in the study are now all freely available online at the Morphosource project “Fischer et al 2022 Proc B – marine amniote teeth”, project ID: 000435369

Arguably one of the most interesting aspects of the study was the simplicity of the results – something which came as no surprise to lead author Valentin Fischer: “I do not like to overhype science and eureka moments… so I’ll be fully honest and say that I somehow expected most of the results we obtained!“. The repeated evolution of simplistic cone-shaped teeth has been noted in previous qualitative and quantitative studies of marine amniotes, but this has never truly been demonstrated across a large sample of clades using three-dimensional shape analyses combined with tooth size and surface complexity. “[The results] seemed sensible and logical, which is a calming feeling for someone initially trained in bone anatomy and systematics” continues Valentin, who has long been interested in marine reptile morphology, relationships, and evolution.

Excerpt from Figure 3 of the paper, demonstrating the morphological constraint of the larger teeth in the top left of the screen (x -axis = principal axis describing shape; y-axis = size).

“Our hope is to help pushing forward the field of marine amniote palaeoecology with explicit incorporation of tooth size and orientation, within a quantitative canvas.” muses Valentin, referring to the methodology put forward in the paper. “This [paper] somehow solves a long-standing conundrum, where animals with similarly shaped teeth were found with clearly distinct gut contents“. One of the great benefits of high-density three-dimensional geometric morphometric approaches is the ability to differentiate very subtle differences between superficially similar objects – assuming one has good resolution data which are not overly damaged! The acquasition of such a sample in a global pandemic was (understandably) problematic, and when asked what the most challenging aspect of the study was, Valentin replied in no uncertain terms: “Getting enough 3D models during COVID, really!“

Finally, I asked Valentin what he sees as the next steps for using this method within the EDDy Lab. While he is very keen for other labs, independent researchers and alike to all try out this method on their own groups of organisms, Valentin certainly has some ideas which he is keen to explore, based around his own research and ongoing projects. “Analysing shape disparity in Iguanodon thumb spikes [as part of the multidisciplinary Iguanodon 2.0 Heritage Science project], belemnite guards, and shark teeth” – watch this space for more fascinating (and high-density) research coming from Liège and collaborating institutions!

The enigmatic “whorl-toothed shark” Helicoprion attacking small cephalopods (which could be belemnites!).
Image by Steve White.

The article is available now online from Proceedings B website, or contact myself or any of the other authors on ResearchGate (e.g. me!) to access it if you cannot get it through your university or institution. A lot of hard work went into the generation and streamlining of the method, and we are all really proud that it could be applied to such a fascinating group of animals.
We all look forward to seeing what the rest of the scientific community does with the method, and what intriguing questions can be answered!

[cover image by A. Gennari – support palaeoartists!]

Mosasaur ecomorphology at world’s end

Finally! After finishing my contract almost a year ago, I can at least reveal the first work I spearheaded as part of the SEASCAPE project at the Université de Liège EDDy Lab. Although it has been a long time coming, I remain very happy and proud of this work, which combines morphometrics, 3D modelling, and disparity – all hallmarks of the EDDy Lab. In our latest contribution, we investigated the final ~15 million years of mosasaur evolution, exploring cranial functional morphology across local and global scales.

Quick recap: mosasaurs were the dominant marine reptiles during the final part of the Cretaceous (83-66 million years ago). [I digress – technically we should call them mosasaurids, but we will keep it at mosasaurs for the purposes of this blog] Most closely related to snakes and lizards (within the order Squamata), mosasaurs first arose in the Cenomanian (c.95Mya), but it was not until the Turonian/Santonian that they truly rose to dominance and occupied numerous trophic levels in Mesozoic marine ecosystems. Examples include the eponymous taxon Mosasaurus – made famous by George Cuvier in the early 1800s before being made way too big in Jurassic World franchise(!) – and Tylosaurus, the “ram-nose” mosasaur, which in all likelihood did NOT have a soft-tissue crest all along it’s back as many of my prehistoric animal books would suggest as a kid! Along with Platecarpus, another one of the earliest marine reptiles I remember learning about when I was a child, I was to meet all these beasts face-to-face many years later as part of the data collection for a study on mosasaur functional morphology.

Paleoartistic representations of arguably the most famous mosasaurs (certainly from my childhood!).
From left: Mosasaurus (©Jonathan Kuo); Tylosaurus (©Henry Sharpe); and Platecarpus (©Henry Sharpe); 3D models rendered from scans taken during this study

Now, most of us have heard about the extinction which wiped out the non-avian dinosaurs 66 million years ago; however, few among us know how the morphological diversity of life was changing prior to that event in non-dinosaurian groups. For example, the diversity and variation in the dominant marine reptiles of the time – the mosasaurs – has only recently started to come to light, first in a feeding and locomotion study from my former university in Bristol (see Cross et al. 2022), and most recently in an in-depth study by myself and colleagues at the Université de Liège, out now in the Proceedings of the Royal Society B: Biological Sciences.

Our aim was to establish how ecomorphologically diverse mosasaurs were when they died out at the end of the Cretaceous, and whether we could explain patterns of diversity or disparity (i.e. diversity in form/function) in the lead-up to the end-Cretaceous mass extinction. To do this, we used a series of high-definition laser and structured-light scans to build 3D models of the skulls of mosasaurs, which we then supplemented with images and measurements from literature sources and first hand photos from “pancaked” specimens. [yes, that’s the term we have coined for specimens which have been fossilised in a near-flat condition!]. Sometimes, I was able to scan pancaked specimens and then reconstruct a life-like skull from that flattened fossil; several of these recontructions were made in order for us to maximise the amount of data we could get from all the different species.

Here is a snapshot of a former work-in-progress where I took a scan from a pancaked mosasaur (Plesioplatecarpus planifrons FHSM VP 2116) and reconstructed it in a much more realistic, life-life position. The original tweet can be found here; note the ongoing discussion on the placement and orientation of the quadrates in these animals!

Within the study, we recorded a series of linear measurements on complete and near-complete skull and dental material from an unprecedented sample of mosasaurs (93 specimens covering 56 species). These linear measurements were used to calculate ratios (often called “functional traits”) which pertain to specific mechanical or sensory outcomes. For example, the length of the mandible from the anterior extremity (i.e. the front of the snout) to the articulation with the cranium can be considered the out-lever for calculating the mechanical advantage for jaw adduction. A similar measurement from the mandible-cranium articulation to the centre of adductor muscle insertion (in this case, the centre of the coronoid process of the mandible) represents the in-lever for mechanical advantage – the ratio of the in-lever to out-lever thus provides a biomechanically-rooted characterisation of jaw addution (i.e. biting). Traits we incorporated included the mechanical advantage of jaw closing and opening; the aspect ratio of the buccal cavity (representing a proxy for the volume of water displaced during jaw addution); the relative depth of the dentary in the middle of the toothrow (representing a proxy for the flexural stiffness of the part of the jaw involved in food processing); and the curvature of the teeth. All combined with characterisations of certain sensory features (e.g. relative orbit size; narial retraction; relative size of the pinneal foramen; etc.), we were able to generate a suite of 16 features which described the ecomorphology of the mosasaur skull – this data could then be processed to generate ecomorphological disparity estimates for discrete time bins and geographical regions.

Examples of measurements taken on a lateral image of Prognathodon solvayi (IRSNB R 33b); image taken from the paper

During our study, we were able to demonstrate (far beyond reasonable doubt) that the last mosasaurs i.e. those present in the Late Maastrichtian (c.70-66 Mya) actually exhibited a different range of ecomorphological characteristics by comparison to those that preceded them. The mosasaurs in the Late Maastrichtian presented a different ‘macroevolutionary landscape’ topology; that is to say, the density of phenotypes (i.e. accumulation of species within a given ecomorphospace) exhibited a very different pattern in the species present in the last throws of the Mesozoic. This ‘landscape’ indicates that mosasaurs in the Late Maastrichtian predominately fall into one of two broad ecomorphological categories:
MEGAPREDATORS – generally large species with deep jaws for resisting bending during biting, sturdy teeth, and relatively large areas for jaw adductor [closing] muscles to produce powerful bites!
GRASPERS – though personally I prefer the accurate but slightly ambiguous term “SUPER-SENSORS“! – small-to-medium sized species with strongly curved teeth, relatively large orbits, large pineal foramina [= large ‘third eye’], relatively slender jaws, and a large tympanic fossae in the quadrate.

Imagine an island of commonly exhibited phenotypes in a sea of non-existant (or very rarely expressed) morphologies. I realise that this, conceptually, is quite hard to imagine!
Now check out the figure above and try again. The largest peak – the light green island – represents the ‘megapredatory’ ecomorphological phenotype just before the K/Pg mass extinction – there is a high density of species with similar functional morphology, represented in the figure by the European tylosaur Tylosaurus bernardi (based on IRSNB R 23). The smaller peak – the slightly darker green “island” – represents the species exhibiting ‘grasping’ ecomorphological traits, with their accompanying suite of sensory characteristics. The example of such a species is Plioplatecarpus (based on TATE V0087).
Figure adapted with permission from MacLaren et al. 2022; © Jamie MacLaren

These two main groups present in the Late Maastrichtian consisted of a phylogenetically diverse assembly of mosasaur species, indicating repeated evolution of similar ecomorphologies in divergent lineages – essentially, multiple different groups undergoing similar selective pressures (maintaining hydrodynamic efficiency, capturing suitable prey, etc.) came to the same functional “conclusions” by way of their cranial morphology. In some cases, in different geographical areas, the variation in functional morphologies exhibited at any one time indicated large-scale shifts in mosasaur community composition – this is where our novel population ecology approach really helped us to see what was going on at global and regional scales

Without getting too much into the maths of the method, our approach to assessing ecomorphological disparity was inspired by the principles of biodiversity quantification. Allow me to explain! In its simplest form, the diversity a localised population within a broader community can be described by its “alpha-diversity” – a count of the species you find in the population (e.g. number of different species of fish in each of the Great Lakes). If we take a step back, look at diversity across a larger geographic scale, we are presented with the “gamma-diversity” for the group (e.g. number of species of fish in The Great Lakes as a whole). Now if we want to understand the diversity of local populations in relation to the larger population, for the purpose of quantifying the differentiation of diversity, we would divide the gamma-diversity value by the mean-average alpha-diversity value across all the populations (e.g. all the individual Great Lakes) to give a value for “beta-diversity“.
Taking these principles for establishing differentiation, we took the approach and applied it to mean disparity values (as calculated from 1000 bootstrap replications), providing us with values or alpha-, gamma-, and beta-disparity respectively. [I like to call them collectively “Greek disparity”, for rather obvious reasons!]. For any disparity queries, all the details are in the paper and the supplement – but rest assured we took the precaution of performing disparity calculations with multiple different disparity metrics (sum of variance, pairwise dissimilarity etc.), and we recovered the same results for all of them, which was reassuring and nice!

Here’s a simple schematic of how population diversity (and disparity) can be partitioned and calculated.
In this example, three different time bins (Time 1, Time 2, Time 3) are presented with arbitrarily generated data for each of the populations. Hopefully this give a quick and simple overview of how “Greek disparity” was calculated in this paper.

Using our “Greek disparity” approach, we were able to demonstrate that mosasaur communities in the Campanian and Maastrichtian were not geographically similar, and by the Late Maastrichtian communities were either ecomorphologically diverse and stable, or declining and becoming more ecologically homogeneous. At the risk of giving the whole game away, I won’t divulge all the results here, but I will provide you with our concluding thoughts. Using this novel combination of population ecology, palaeobiology and functional morphology, we highlight how global and regional patterns of mosasaur ecomorphology may have come about and relate these fluctuations to patterns in other groups of marine reptiles. Moreover, our analysis of mosasaur functional morphology tells a cautionary tale about taking global diversity (and disparity) on face value in the lead-up to mass extinction events.

I speak for all the paper authors when I say thank you for reading through this blog – hopefully it has offered more info on this new study and has intrigued you on the biology of mosasaurs. Find the paper here, the 3D models we scanned can be viewed on our Morphosource project here, and look out for more mosasaur-related content coming from the University of Liège EDDy Lab very soon! Until then, Hakuna Squamata!

Changing the face of Brachyodus

I will be the first to admit that I know very little about a lot of ungulate groups! Performing a PhD on one of the most enigmatic groups of hooved mammals (the tapirs) did not prepare me for the myriad of different forms of hooved mammals which lived before, alongside, and after many of my study animals. Thus, when it came to identifying a mysterious skull which was sat on the bottom shelf of a display cabinet in the Evoution and Diversity Dynamics lab of the University of Liege, I was a little stumped. Our older records initially suggested “a rhino”, and some more recent records stated “Anthracotherium“. In fairness, the latter identification was much closer to the mark! Little did I know at the time, but the skull which had been neglected for years and left at the bottom of the pile turns out to be a very special find.

Mystery skull moved out of the shadows and onto the table in preparation for laser scanning

Having established that this skull was pretty cool (look at those chompers!!), I set to a literature search to try a preliminary identification. Cranium-wise, I couldn’t find anything similar, and at the time my assumption was that the cranium and mandible are from the same individual. I had more luck with the mandible, which bore a striking resemblence to the near-complete mandible of Brachyodus depereti (referred to here as ‘Masritherium’ depereti after Fortau 1918). After consultation with the authors of Miller et al. (2014), referal to Brachyodus (sp. indet.) was favoured, and at that point I put little thought to the matter. The important thing was that we had a cool anthracothere in the collection, and as a very well preserved 3D specimen it seemed a shame not to laser scan it! So that’s exactly what I did, and we uploaded the scan to the lab Sketchfab site for the world to see.

Aligned scans of Brachyodus mandible (left) and cranium (right), laser scanned with HandyScan^TM 300 ©Jamie A. MacLaren. Each scan is taken and aligned with one another – the different colours in the images represent a separate scan which was then aligned in VXElements^TM software

Enter Dr. Martin Pickford. As a decorated veteran of vertebrate palaeontology across Europe and Africa, Martin has been at the vanguard of anthracothere palaeobiology, palaeoecology and phylogenetic affinity. Much has been made of the potential links between anthracotheres (such as Brachyodus) and the hippopotamids (e.g. modern river hippo Hippopotamus and pygmy hippo Choeropsis). Some have argued that anthracotheres exhibit sufficiently similar cranial and post-cranial features to hippopotamids to indicate that, at the very least, anthracotheres were habitually semi-aquatic (e.g. Orliac et al. 2013), and may have been ancestral to the later hippopotamid lineages. For me, as a functional anatomist, the phylogenetic aspects of anthracotheriids vs. hippopotamids are not of greatest interest or importance; HOWEVER, I do respect that this is a question which raises blood pressures and is very much in need of resolution. Martin and I therefore hope that our new contribution (and associated 3D data) will be used responsibly in the pursuit of phylogenetic resolution for Brachyodus, and Anthracotheriidae in general. And so – to our contribution!

Adaptation of figure from Orliac et al. (2013) depicting Hippopotamus (top) and Brachyodus (bottom). In this reconstruction (based on the fossil MNHT SAF 001; bottom right), the Brachyodus appears very hippo-like, with eyes protruding above the waterline. However, this reconstruction was made without any knowledge of the maxillary/premaxillary region of the skull – something which our new specimen provides.

Our critical description of the new specimen from the collections of the University of Liège (ULg) flagged up some very interesting insights into the skull morphology and composition of the genus Brachyodus. For starters, the chances are very high that the cranium and the mandible of the ULg specimen are in fact not from the same individual – this will need rigerous testing, but as far as tooth-row measurements are concerned, it is highly indicative of two separate specimens. That said, our interpretation is that they do belong to the same species – sadly (for me at least!), that species is already known, so we are not able to describe a new species of the genus. However, the specimens does represent the best known remains of the species Brachyodus onoideus. The diagnosis to this species was made based on dental metrics, in addition to general size and appearence, plus additional key features hitherto unknown for the species. Due to the likelihood that the specimens of mandible and cranium belong to separate individuals, we re-assigned the mandible the specimen number ULg M 5000a (as it is the only part of the specimen labelled with a specimen number in the first place!!), and the cranium is now assigned as ULg M 5000b. 3D models of both specimens can be found on the Morphosource.org online repository, just look for the project ID: 000421549 (“Pickford & MacLaren (2022) Brachyodus onoideus”). The provenance of the specimens is not accurately known – very little information was available for the skull beyond the specimen number on the mandible. Our assessment is that these specimens share colouration, preservation and match the species presence with fossils from the Sables de l’Orléanais from localities in France (Nancray, Chilleurs-aux-Bois); we therefore comfortably attribute the specimens geographically to south-western Europe, potentially central France, during the Early Miocene.

An example of specimens from the Sables de l’Orléanais, figured in Gagnaison (2020), including Brachyodus onoideus (bottom right)

From our description of the new material, we offer two main conclusions which – although in need of more rigorous assessment – provide intriguing insights into the biology of Brachyodus, and potentially anthracotheres at large.

1) The “Semi-Aquatic” Conundrum
A number of researchers have previously suggested anthracotheres are closely related to hippopotamids, and that elements of the skull are indicative of a habitual semi-aquatic ecology in these animals. Our finding of the skull, with its comparatively low orbits (eye-sockets) with no obvious dorsal protrusion – as is seen in Hippopotamus, a high and sharp sagittal crest, and a steeply oriented occipital condyle (articulation between the skull and the atlas, or first vertebra). Let us examine these in turn:

The orbits being placed near the centre of the skull is actually a very common occurence for a number of ungulates, especially those with higher sagittal crests and those which are not habitual grazers. Grazing animals like horses or sheep have elevated orbits, enabling them to retain a good view of their surroundings while feeding with their mouths close to the ground. In Hippopotamus, the orbits are in fact nested within protrusions from the skull, enabling them to peer above the water-line without raising their entire head. As we demonstrate in our study, the orbits of Brachyodus are not anywhere near as elevated as in Hippopotamus – that said, the pygmy hippo Choeropsis does not exhibit protruding orbits, though its eyes are elevated on the skull. The absence of this adaptation to a habitual semi-aquatic existence in the ULg specimens of Brachyodus onoideus certainly suggests that this species / genus was not capable of raising it’s eyes above the water without revealing the majority of the rest of its head!

The high and narrow sagittal crest present in the ULg M 5000b cranium of Brachyodus is raised high above the orbits, and is indicative of quite substantial temporalis musculature in this species. Similarly high crests are observed in semi-aquatic ungulates alive today, such as tapirs; however, possibly the most similar crest morphology among modern ungulates is that of the (very definitely not semi-aquatic) Old World camelids (e.g. Camelus dromedarius, the dromedary)! The height of the sagittal crest does not preclude Brachyodus from being semi-aquatic – many tapirs with pronounced crests spend much of their day in water – but the combination of high crest and low orbits do not lend support to a “Hippopotamus-like” ecology for Brachyodus. Our findings from this study are further supported by a 2021 microanatomical analysis by Houssaye et al. of the long bones of hippos and other ungulates, suggesting that Brachyodus onoideus was maybe, probably, slightly semi-acquatic(!), likely to have immersed itself on occassion, and presumably putting it in a similar ecological category with moose Alces, lechwe / waterbuck Kobus spp., and tapirs Tapirus.

Comparison of skull of Brachyodus onoideus (ULg M 5000a/b) with hippopotamids (top), camelids (middle) and tapirids (bottom).
Hippopotamids: Choeropsis liberiensis (Uwe Gille, Halle (Saale), Germany); Hippopotamus amphibius (DUNUC 1990; based on 3D model available on Sketchfab at https://sketchfab.com/uod_museums).
Camelids: Camelus dromedarius (NHMB 2128; Martini et al. 2018); Tanymykter longirostris (CM 2498; Peterson 1911; Lynch et al. 2020). Tapirids: Tapirus pinchaque (MNHN 1982-34); Tapirus veroensis (UF/FGS 221).
Black line passes through the centre of the Brachyodus orbit, with the two grey lines marking the dorsal extent of the sagittal crest and the margin of the alveoli of the maxillary toothrow – clearly, both hippopotamids’ orbits are positioned well above the black line marked by the orbit of Brachyodus onoideus.

2) How the head was held
As for the inclination of the occipital condyle, this feature is indicative of how the skull is carried on the neck. From previous studies (e.g. Pickford 2015) it has been shown that Brachyodus did not have a short neck like that of hippopotamids and suids, but rather posessed a relatively long neck like many modern-day ruminants (incl. cervids & bovids). The orientation of the occipital condyle in the ULg M 5000b cranium also lends support to the head being held above the origin of the cervical column (i.e. neck vertebrae), and thus may have looked not too dissimilar to a horse or okapi in the positioning of its head and neck on the body.
THAT IS NOT TO SAY that Brachyodus is in any way phylogenetically closer to equids or giraffids than to suoids or hippopotamids!! Simply the carriage of the skull on the neck seems more indicative of ruminant / equine ungulate morphologies than suoid / hippopotamid ungulate morphologies.

Highly speculative reconstruction of cranial appearance of Brachyodus onoideus
Reconstruction by Orliac et al. (2013) based on crushed skull of MNHT SAF 001 (top), compared to life-restoration by MacLaren (this blog) based on 3D preserved skull of ULg M 5000a/b (bottom). Note the continuing lack of knowledge regarding the orientation of the nostrils, but also the clear distinction between the dorsal skull morphology and the region occupied by the eye. ©Jamie A. MacLaren

In this study, we justifiably fall short of assigning any phylogenetic significance to our interpretations based on ULg M 5000a/b – as they are functional and ecological in nature – but we do strongly encourage researchers working on establishing phylogenetic relationships within the Anthracotheriidae or within (Cet)Artiodactyla in general to inspect the specimens at the University of Liège collection, as they may be pivotal in understanding relationships within this much-maligned and hotly debated clade. We hope the reader has wheted their appetite for more info on the Liège specimens – the article can be found in a special issue of Historical Biology commemorating the work of Prof. Jorge Morales, and also a my ResearchGate research page. Spread the word – Brachyodus was almost certainly probably not semi-aquatic in the way researchers used to think!